Research And Implementation Of The Water Level Monitoring System Based On Image Recognition

Currently,water security and water resources protection have become one of the most important factors affecting our social,economic and ecological development.There are many ways to measure the water level,with the help of kinds of water gauges.However,those water level gauges are confronted with several problems: complexity of work,vulnerability to environmental factors,and unsuitability in long distance measurements.Therefore,many domestic hydrological stations still apply manual method to monitor the water level,which may pose severe threats towards the safety of monitors.Moreover,the level measured with this method is neither accurate nor real-time.It has become an inevitable trend that the traditional water level detection by manual work be replaced by intelligent monitoring system.In order to overcome the above shortcomings,we design a remote water level monitoring system by using image recognition technology and wireless communication technology.The main contents of this thesis are as follows:Several water level recognition algorithms are studied,and a water level recognition method based on dictionary learning is proposed.The overall process of water level monitoring is planned and designed for a specific environment.This thesis designs an intelligent water regimen monitoring system,which consist of remote monitor station and monitoring center.The remote monitor station contains an embedded data acquisition module in order to collect environmental data of water level and image periodically.The collected data is then sent to a server via cellular network,where it is stored in a database and processed to determine early warning of flood in the area based on historical data.The designed system not only solves measuring accuracy problem,but also increases working efficiency greatly.The automatic water level detection technology proposed in this paper has been tested in the Nanjing River section monitoring system.The system is tested to run well and the measurement error is within 5cm,which shows stable performance in practical operation.